@cryer/star-cli 0.3.8 → 0.3.10

package/README.md

@@ -66,7 +66,7 @@ maxSteps = 100
 contextCompaction = "summary"     # summary | truncate — how over-budget history is compacted
 streamMaxRetries = 5              # extra attempts per model request on transient failures
                                   # (network error, 429/5xx, idle cutoff, empty reply); 0 disables
-streamIdleTimeoutSec = 20         # seconds with no stream output before a stalled response is
+streamIdleTimeoutSec = 60         # seconds with no stream output before a stalled response is
                                   # ended gracefully (some relays never close the stream)
 streamFirstChunkTimeoutSec = 300  # seconds to wait for the very first content part — slow
                                   # thinking-model endpoints can stay silent for minutes
@@ -116,6 +116,11 @@ model = "gpt-4o"                  # model id sent to the API
                                   # provider (low/medium/high are common; some add minimal/max...).
                                   # Unset = server default. Ignored for anthropic. The /model
                                   # picker also offers these levels after a model pick.
+# vision = false                  # mark a text-only model: read_image/screenshot then decline with
+                                  # a text error instead of attaching images the endpoint would
+                                  # reject. Unset = images allowed. (Without it, a 4xx rejecting
+                                  # images outright still self-heals: the images are stripped from
+                                  # the history and the request retried once.)
 # streamIdleTimeoutSec = 60       # per-model overrides of the global stream watchdog timeouts
 # streamFirstChunkTimeoutSec = 600
 # optional per-model pricing in USD per 1M tokens — enables the $ estimate in /cost and /usage.
@@ -172,7 +177,7 @@ model = "kimi-for-coding"
 temperature = 1        # mandatory here — the endpoint rejects any other value
 contextMaxTokens = 128000
 # raise these when the relay buffers long generations:
-# streamIdleTimeoutSec = 60        # default 20
+# streamIdleTimeoutSec = 120       # default 60
 # streamFirstChunkTimeoutSec = 600 # default 300
 ```
 
@@ -384,3 +389,12 @@ pnpm test:pipeline   # layered pipeline incl. optional live LLM smoke (needs an
 CI runs lint, typecheck, tests, and build on Ubuntu + Windows against Node 20/22 (`.github/workflows/ci.yml`).
 
 Architecture: `src/cli` (Ink UI), `src/agent` (main loop), `src/llm` (Vercel AI SDK provider layer), `src/tools`, `src/context` (token budget + compaction), `src/permissions`, `src/session`, `src/config`.
+
+## Benchmark
+
+[Terminal-Bench 2.0](https://github.com/laude-institute/terminal-bench), 20 randomly sampled tasks, same model for both agents:
+
+| Agent | Model | Pass rate | Avg tokens | Avg duration |
+|---|---|---|---|---|
+| Star CLI 0.3.9 | Kimi K3 | 17/20 (85%) | 129.2K | 414s |
+| terminus-2 2.0.0 | Kimi K3 | 9/20 (45%) | 49.5K | 484s |

package/README.zh-CN.md

@@ -65,7 +65,7 @@ maxSteps = 100
 contextCompaction = "summary"     # summary | truncate —— 历史超出预算时的压缩方式
 streamMaxRetries = 5              # 单次模型请求在瞬时失败（网络错误、429/5xx、idle 截断、
                                   # 空回复）时的额外重试次数；0 表示禁用
-streamIdleTimeoutSec = 20         # 流没有任何输出的秒数，超过后将停滞的响应优雅结束
+streamIdleTimeoutSec = 60         # 流没有任何输出的秒数，超过后将停滞的响应优雅结束
                                   #（有些中转站从不关闭流）
 streamFirstChunkTimeoutSec = 300  # 等待第一个内容分片的秒数——慢速思考模型端点在接受请求后
                                   # 可能静默数分钟
@@ -114,6 +114,10 @@ model = "gpt-4o"                  # 发送给 API 的模型 id
                                   # minimal/max 等），因此接受任意字符串，由服务端校验。
                                   # 不设置 = 服务端默认。anthropic 协议下忽略。
                                   # /model 选完模型后也会接着让你选择推理强度。
+# vision = false                  # 标记纯语言模型：read_image/screenshot 会直接返回文本错误，
+                                  # 不再把图片塞进历史导致端点报 4xx。不设置 = 允许图片。（即使
+                                  # 不设置，端点明确拒绝图片的 4xx 也会自愈：剥掉历史中的图片后
+                                  # 自动重试一次。）
 # streamIdleTimeoutSec = 60       # 按模型覆盖全局的流看门狗超时
 # streamFirstChunkTimeoutSec = 600
 # 可选的按模型定价（美元/1M token）——启用 /cost 和 /usage 中的美元估算。
@@ -168,7 +172,7 @@ model = "kimi-for-coding"
 temperature = 1        # 此处必须设置——该端点拒绝任何其他取值
 contextMaxTokens = 128000
 # 中转站长时间缓冲整段生成时可调大：
-# streamIdleTimeoutSec = 60        # 默认 20
+# streamIdleTimeoutSec = 120       # 默认 60
 # streamFirstChunkTimeoutSec = 600 # 默认 300
 ```
 
@@ -380,3 +384,12 @@ pnpm test:pipeline   # layered pipeline incl. optional live LLM smoke (needs an
 CI 在 Ubuntu + Windows 上、Node 20/22 下运行 lint、typecheck、测试和构建（`.github/workflows/ci.yml`）。
 
 架构：`src/cli`（Ink UI）、`src/agent`（主循环）、`src/llm`（Vercel AI SDK 服务商层）、`src/tools`、`src/context`（token 预算 + 压缩）、`src/permissions`、`src/session`、`src/config`。
+
+## 基准测试
+
+[Terminal-Bench 2.0](https://github.com/laude-institute/terminal-bench)，随机抽取 20 个任务，两个 agent 使用相同模型：
+
+| Agent | 模型 | 通过率 | 平均 token 消耗 | 平均耗时 |
+|---|---|---|---|---|
+| Star CLI 0.3.9 | Kimi K3 | 17/20 (85%) | 129.2K | 414s |
+| terminus-2 2.0.0 | Kimi K3 | 9/20 (45%) | 49.5K | 484s |

package/dist/chunk-4IRY44TG.js

@@ -0,0 +1,34 @@
+#!/usr/bin/env node
+
+// src/llm/registry.ts
+function listModels(config) {
+  return config.models;
+}
+function resolveModelConfig(config, modelName) {
+  const name = modelName ?? config.defaultModel;
+  const model = config.models.find((m) => m.name === name);
+  if (!model) {
+    const available = config.models.map((m) => m.name).join(", ") || "(none)";
+    throw new Error(`Model "${name}" not found. Available models: ${available}`);
+  }
+  return model;
+}
+function resolveStartupModel(config, sessionModel) {
+  if (sessionModel) {
+    if (config.models.some((m) => m.name === sessionModel)) {
+      return { name: sessionModel, notice: null };
+    }
+    return {
+      name: config.defaultModel,
+      notice: `Session model "${sessionModel}" not found in config \u2014 using default model.`
+    };
+  }
+  return { name: config.defaultModel, notice: null };
+}
+
+export {
+  listModels,
+  resolveModelConfig,
+  resolveStartupModel
+};
+//# sourceMappingURL=chunk-4IRY44TG.js.map

package/dist/chunk-4IRY44TG.js.map

@@ -0,0 +1 @@
+{"version":3,"sources":["../src/llm/registry.ts"],"sourcesContent":["import type { ModelConfig, StarConfig } from \"../config/schema\";\n\nexport function listModels(config: StarConfig): ModelConfig[] {\n  return config.models;\n}\n\nexport function resolveModelConfig(config: StarConfig, modelName?: string): ModelConfig {\n  const name = modelName ?? config.defaultModel;\n  const model = config.models.find((m) => m.name === name);\n  if (!model) {\n    const available = config.models.map((m) => m.name).join(\", \") || \"(none)\";\n    throw new Error(`Model \"${name}\" not found. Available models: ${available}`);\n  }\n  return model;\n}\n\n// Pick the model at startup. An explicit --model already sits in\n// config.defaultModel (loader override), so it wins naturally; otherwise a\n// resumed session's recorded model is restored when it still exists in the\n// config. A stale recorded model falls back to the default with a notice.\nexport function resolveStartupModel(\n  config: StarConfig,\n  sessionModel?: string | null,\n): { name: string | undefined; notice: string | null } {\n  if (sessionModel) {\n    if (config.models.some((m) => m.name === sessionModel)) {\n      return { name: sessionModel, notice: null };\n    }\n    return {\n      name: config.defaultModel,\n      notice: `Session model \"${sessionModel}\" not found in config — using default model.`,\n    };\n  }\n  return { name: config.defaultModel, notice: null };\n}\n"],"mappings":";;;AAEO,SAAS,WAAW,QAAmC;AAC5D,SAAO,OAAO;AAChB;AAEO,SAAS,mBAAmB,QAAoB,WAAiC;AACtF,QAAM,OAAO,aAAa,OAAO;AACjC,QAAM,QAAQ,OAAO,OAAO,KAAK,CAAC,MAAM,EAAE,SAAS,IAAI;AACvD,MAAI,CAAC,OAAO;AACV,UAAM,YAAY,OAAO,OAAO,IAAI,CAAC,MAAM,EAAE,IAAI,EAAE,KAAK,IAAI,KAAK;AACjE,UAAM,IAAI,MAAM,UAAU,IAAI,kCAAkC,SAAS,EAAE;AAAA,EAC7E;AACA,SAAO;AACT;AAMO,SAAS,oBACd,QACA,cACqD;AACrD,MAAI,cAAc;AAChB,QAAI,OAAO,OAAO,KAAK,CAAC,MAAM,EAAE,SAAS,YAAY,GAAG;AACtD,aAAO,EAAE,MAAM,cAAc,QAAQ,KAAK;AAAA,IAC5C;AACA,WAAO;AAAA,MACL,MAAM,OAAO;AAAA,MACb,QAAQ,kBAAkB,YAAY;AAAA,IACxC;AAAA,EACF;AACA,SAAO,EAAE,MAAM,OAAO,cAAc,QAAQ,KAAK;AACnD;","names":[]}